Many people are unable to speak, either from birth or due to something that happened later in life. Thanks to modern technology, we now have voice synthesizers that enable them to translate their written words into sounds. Stephen Hawking is perhaps the person whom we are most familiar with who uses this.
But apparently there are only a handful of synthetic voices available, with the result that the speech of the person may not match their age, gender, physique, region of origin, ethnicity, and the like, making for some incongruity. Hawking, for example, speaks with an American accent even though he is English.
Via reader Norm, I heard about a project that takes the voice synthesizing technology that we have to create more authentic sounds. It does this is by taking whatever sounds that the speech-impaired person makes, however incoherent, and combining it with the basic elements of sounds extracted from voices that have been ‘donated’ by others that are close to them in the desired characteristics to generate personalized synthetic voices, so that the voiceless person feels like it really belongs to them.
The process is similar to the way that Siri’s voice is generated. In order to do this, the scientists need a data bank of voices to work with and if you are willing to spend about three hours recording your voice, you can sign up to be a voice donor here. Norm has already done so and so have I.
In this video, scientist Rupal Patel describes how the system works and the results they have obtained. It is quite moving.